Parameterization and Statistical Analysis of Hurricane Waves

Abstract

Recently, Gulf coast communities have experienced significant damage from landfalling hurricanes. While the effects of hurricane surge on coastal communities have been examined and better defined, risk of damage due to hurricane waves is less quantified. This thesis presents the Wave Response Function (WRF) methodology. Hurricanes are parameterized in the form of non-dimensional equations incorporating key physical hurricane parameters. The non-dimensional equations are then combined with a fully developed sea state cap (Young and Verhagan 1996) to form the open coast and bay methodologies. This approach yields root mean square errors (RMSE) ranging from 0.01-0.46 m, with the majority of points below 0.3 m. This approach yields small bias values. The WRF method was compared to Hurricane Ike data (Kennedy et al. 2011) and yielded RMSE of 0.67 meters despite the higher depths of the recording stations. The WRF method was also compared to Taylor’s (2012) parameterized wave equations, with mean RMSE improvements ranging from 0.13-0.32 m.

Once WRF coefficients are adjust to minimize RMSE at each station under consideration, extreme value analysis via the Joint Probability Method with Optimal Sampling (JPM-OS) was conducted. When applied to Panama City, FL the JPM-OS methodology yielded extreme value statistics for 179 stations of interest. Maps detailing the spatial extents of the 100 and 1000 year maximum wave event were created using ArcGIS.